I built a SOAP web service using WCF as a part of a demo. However, as it turned out, the client software I was given can only handle HTTP as transport (the WCF service I did accepts requests directly from TCP payload). So I had to convert it into a RESTful web services. I have basicly a couple of options to implement the RESTful web services:
- Modify the existing WCF project to support REST style according to MSDN
- Create a ASP.NET project and reuse the code that I created in the WCF project.
- Create a Java web project and rewrite the code.
The 1st option seems the best choice on the surface. But there are still too much of hand-rolling of boiler-plate type of code to be done in option 1. You have to process the HTTP request method using if/switch ... yuck.
The 2nd option will save me some time because I can reuse the LINQ to XML code. But the code is very short, so the saving is not big. Again, the disadvantage of this option is that I have to create all the virtual path mapping (extending the VirtualPathProvider class) myself. The time to spend on building the ASP.NET web application from scratch will outweigh the savings on reusing the bit of XML query code (although there is a pretty good sample by Doug Seven from where I can steal some code.)
NetBeans (v6.1) has a RESTful web services (JSR-311) plugin using Jersey libraries. The wizard can create all the boilerplate codes and configurations required to get a RESTful service up and running by a few simple steps. The downside of this approach is that I have to rewrite the web service implementation methods - XML query, HTML scraping... However, the code for those is quite simple and should be easy to rewrite.
So, ASP.NET is lagging behind in the area of RESTful web services. I guess to get the similar features offered by JSR-311 and NetBeans, one has to wait for the upcoming MVC Framework for ASP.NET. An excellent tutorial and example project using the MVC Framework can be found here.
Since time is of essense in this exercise, I chose option 3.
I started by creating a Java Web Application (called SvdemoRestful) in NetBeans. Then adding a RESTful web service is simply right-clicking the project and choose New -> RESTful Web Services from Patterns... NetBeans also offers a wizard to create RESTful Web Services from Entity Class. I did not need to use Entity Class because I am going to use XPath to query a simple XML file (as my database) and return the same XML snippet to the service consumer (i.e. I don't even need to convert the XML into a Java bean);and I don't want to use the ORM tools and create Persistent Objects. Then it's just a matter of filling in the self-explanatory fields in the wizard.
Once this is done, the skeleton code is generated along with all the necessary web configuration settings (e.g. url mapping). I want to have my RESTful service to have URI like below so that the GET method will return the first promotion that matches the keywords (actually, it is better to use query parameters instead of path parameters for keyword searching services):
http://hostname/SvdemoRestful/resources/promo/jazz, action
To achieve this, the @Path annotation is defined as
@Path("promo/{keywords}")
I used XPath to retrieve the Promotion record from the data.xml file. I need to make the keyword matching case in-sensitive. In XPath 2.0 there is an upper-case() function, but it does not seem to be supported by the XPath libraries. So I have to use the following XPath statement:
//Promotion[contains(translate(Tags,'abcdefghijklmnopqrstuvwxyz','ABCDEFGHIJKLMNOPQRSTUVWXYZ'),'keywordsToBeMatched')]
as opposed to
//Promotion[contains(upper-case(Tags),'keywordsToBeMatched')]
So the resulting code for the promo RESTful web service:
/*
* PromoResource
*
*/
package svdemo;
import java.util.logging.Level;
import java.util.logging.Logger;
import javax.ws.rs.core.Context;
import javax.ws.rs.core.UriInfo;
import javax.ws.rs.ProduceMime;
import javax.ws.rs.ConsumeMime;
import javax.ws.rs.PUT;
import javax.ws.rs.Path;
import javax.ws.rs.GET;
import javax.ws.rs.PathParam;
import javax.xml.parsers.DocumentBuilder;
import javax.xml.parsers.DocumentBuilderFactory;
import org.w3c.dom.Node;
import javax.xml.xpath.XPath;
import javax.xml.xpath.XPathConstants;
import javax.xml.xpath.XPathExpression;
import javax.xml.xpath.XPathFactory;
import org.w3c.dom.Document;
import org.w3c.dom.NodeList;
/**
* REST Web Service
*
* @author ROMENL
*/
@Path("promo")
public class PromoResource {
static String EMPTY="";
@Context
private UriInfo context;
/** Creates a new instance of PromoResource */
public PromoResource() {
}
/**
* Retrieves representation of an instance of svdemo.PromoResource
* The URL parameters are:
* keywords - comma delimited words to be matched
* @return an xml string representing the PromoInfo
*/
@GET
@Path("{keywords}")
@ProduceMime("text/xml")
public String getXml(@PathParam("keywords") String keywords ) {
if (keywords == null || keywords.length()==0) {
return EMPTY;
}
String[] words = keywords.split(",");
try {
DocumentBuilderFactory domFactory = DocumentBuilderFactory.newInstance();
domFactory.setNamespaceAware(true);
DocumentBuilder builder = domFactory.newDocumentBuilder();
Document doc = builder.parse("data.xml");
XPathFactory factory = XPathFactory.newInstance();
XPath xpath = factory.newXPath();
for(String keyword : words) {
XPathExpression expr = xpath.compile(
"//Promotion[contains(translate(Tags,'abcdefghijklmnopqrstuvwxyz','ABCDEFGHIJKLMNOPQRSTUVWXYZ'),'"
+keyword.trim().toUpperCase()+"')]");
Node result = (Node)expr.evaluate(doc, XPathConstants.NODE);
if(result!=null) {
StringBuffer sb=new StringBuffer("");
NodeList nodes = result.getChildNodes();
for(int i=0; i<nodes.getLength(); i++) {
Node node = nodes.item(i);
if(!"#text".equals(node.getNodeName())){
sb.append("<"+node.getNodeName()+">");
sb.append(node.getTextContent());
sb.append("");
}
}
sb.append("");
return sb.toString();
}
}
return EMPTY;
} catch (Exception ex) {
Logger.getLogger(PromoResource.class.getName()).log(Level.SEVERE, null, ex);
return EMPTY;
}
}
/**
* PUT method for updating or creating an instance of PromoResource
* @param content representation for the resource
* @return an HTTP response with content of the updated or created resource.
*/
@PUT
@ConsumeMime("application/xml")
public void putXml(String content) {
}
}
In another service I needed to pass in a URL so that the service will get the web page of the URL and scrape the HTML code (yes, scraping HTML is a bad practice, but I don't have any alternatives and it's OK for demo purposes). The HTML scraping is easily handled by using the
javax.swing.text.html.HTMLEditorKit. The problem is that the input parameter to this RESTful service contains forward-slashes ('/') and it will confuse the web server if I give the URLs like so:
http://hostname/SvdemoRestful/resources/webPageKeywords/http://localhost/someWebPage.html
http://hostname/SvdemoRestful/resources/webPageKeywords/'http://localhost/someWebPage.html'
According to JSR-311, the above urls should be specified as @Path("webPageKeywords/{url:.+}"), but it did not work. I tried different values of the
limited and
encode attributes of the @Path(), but to no avail. So I had to resort to passing the parameters as the URL's query parameter:
http://hostname/SvdemoRestful/resources/webPageKeywords?url=http://localhost/someWebPage.html
The corresponding @Path annotation is just @Path("webPageKeywords"). The corresponding service implementation code snippet:
/**
* REST Web Service
*
* @author ROMENL
*/
@Path("webPageKeywords")
public class WebPageKeywordsResource {
@Context
private UriInfo context;
/** Creates a new instance of WebPageKeywordsResource */
public WebPageKeywordsResource () {
}
/**
* Retrieves representation of an instance of svdemo.WebPageKeywordsResource
* @return empty string if no keywords found; otherwise, comma delimited keywords
*/
@GET
@ProduceMime("text/plain")
public String getText(@QueryParam("url") String urlString) {
if(urlString==null || urlString.length()==0)
return "";
try {
URL url = new URL(urlString);
HTMLEditorKit kit = new HTMLEditorKit();
HTMLDocument doc = (HTMLDocument) kit.createDefaultDocument();
doc.putProperty("IgnoreCharsetDirective", Boolean.TRUE);
Reader HTMLReader = new InputStreamReader(url.openConnection().getInputStream());
kit.read(HTMLReader, doc, 0);
String keywords="";
Element e=doc.getElement("ProfileMusic");
if(e!=null) {
keywords=doc.getText(e.getStartOffset(), e.getEndOffset()-e.getStartOffset());
}
e=doc.getElement("ProfileFilms");
if(e!=null) {
keywords+=","+doc.getText(e.getStartOffset(), e.getEndOffset()-e.getStartOffset());
}
return keywords;
} catch (Exception ex) {
Logger.getLogger(WebPageKeywordsResource .class.getName()).log(Level.SEVERE, null, ex);
return "";
}
}
/**
* PUT method for updating or creating an instance of WebPageKeywordsResource
* @param content representation for the resource
* @return an HTTP response with content of the updated or created resource.
*/
@PUT
@ConsumeMime("text/plain")
public void putText(String content) {
}
}
I am looking forward to the new ASP.NET MVC Framework and will definitely rewrite these services using it.
Orion & OC4J
Beijing 2008
